Automation Anywhere made a bold claim at its annual customer conference this week, claiming it has taken a "significant step forward towards Artificial General Intelligence (AGI) for the enterprise" with the launch of what it calls the industry's first Process Reasoning Engine.

The company announced the expansion of its Agentic Process Automation (APA) system, introducing technology that CEO Mihir Shukla says represents "a fundamental reimagining of how work gets done" rather than mere incremental improvement.

At the heart of the announcement is the Process Reasoning Engine (PRE), an AI system designed to understand enterprise context and dynamically orchestrate work across different departments and industries. Unlike general-purpose AI models, the PRE is specifically tailored for business processes in sectors ranging from banking and healthcare to manufacturing and customer service.

"Our PRE is a pivotal advancement with specialized insight into processes across industries and departments," Shukla explained.

"PRE will fundamentally change how entire departments and companies work, enabling them to go from idea-to-action with speed and efficiency."

The company claims the technology delivers tangible results: 3x higher efficiency in building end-to-end automation and 60% greater automation resiliency compared to standalone large language models or traditional automation tools.

Computer-Using AI Agents

The company also announced the introduction of Enterprise UI Agents - that Automation Anywhere claims that can interact with Web-based applications like humans do. These agents can understand context, adapt in realtime, and execute complex, multi-step tasks across dynamic web interfaces.

The expanded system goes beyond individual AI agents to orchestrate what the company calls "long-running, mission-critical processes." The platform can coordinate AI agents, traditional RPA bots, APIs, documents, and human workers across different systems and vendors.

Automation Anywhere is positioning itself for broader industry adoption by supporting emerging interoperability standards, including Google Cloud's Agent-to-Agent protocol and Anthropic's Model Context Protocol. This allows integration with major platforms like AWS Bedrock, Google Agentspace, Microsoft Copilot, and Salesforce Agentforce.

The company claims organizations using the system can automate up to 80% of their operations while maintaining enterprise-grade security and governance features, including PII data masking and AI guardrails.

Maureen Fleming, Program VP at IDC's Worldwide Intelligent Process Automation Market Research, called the announcement "a big moment for enterprise automation efficiency."

"Automation Anywhere has systematically collected data about task execution for several years, and is now able to combine that with intelligent, goal-driven reasoning to speed up the development of agentic capabilities," Fleming noted.

https://www.automationanywhere.com/